Merge branch 'master' into ghc7.0

Utility/Url.hs

@@ -19,6 +19,7 @@ import Control.Monad
 import qualified Network.Browser as Browser
 import Network.HTTP
 import Network.URI
+import Data.Maybe
 
 import Utility.SafeCommand
 import Utility.Path
@@ -87,12 +88,32 @@ get url =
 
 {- Makes a http request of an url. For example, HEAD can be used to
  - check if the url exists, or GET used to get the url content (best for
- - small urls). -}
+ - small urls).
+ -
+ - This does its own redirect following because Browser's is buggy for HEAD
+ - requests.
+ -}
 request :: URI -> RequestMethod -> IO (Response String)
-request url requesttype = Browser.browse $ do
-	Browser.setErrHandler ignore
-	Browser.setOutHandler ignore
-	Browser.setAllowRedirects True
-	snd <$> Browser.request (mkRequest requesttype url :: Request_String)
+request url requesttype = go 5 url
 	where
+		go :: Int -> URI -> IO (Response String)
+		go 0 _ = error "Too many redirects "
+		go n u = do
+			rsp <- Browser.browse $ do
+				Browser.setErrHandler ignore
+				Browser.setOutHandler ignore
+				Browser.setAllowRedirects False
+				snd <$> Browser.request (mkRequest requesttype u :: Request_String)
+			case rspCode rsp of
+				(3,0,x) | x /= 5 -> redir (n - 1) u rsp
+				_ -> return rsp
 		ignore = const $ return ()
+		redir n u rsp = do
+			case retrieveHeaders HdrLocation rsp of
+				[] -> return rsp
+				(Header _ newu:_) ->
+					case parseURIReference newu of
+						Nothing -> return rsp
+						Just newURI -> go n newURI_abs
+							where
+								newURI_abs = fromMaybe newURI (newURI `relativeTo` u)

doc/forum/How_to_expire_old_versions_of_files_that_have_been_edited__63__.mdwn

@@ -0,0 +1,7 @@
+My annex contains several large files that I have unlocked, edited, and committed again, i.e. the annex contains the version history of those files. However, I don't want the history -- keeping the latest version is good enough for me. Running `git annex unused` won't detect those old versions, though, because they aren't unused as old Git revisions still refer to them. So I wonder:
+
+1. What is the best way to get rid of the old versions of files in the annex?
+
+2. What is the best way to detect old versions of files in the annex?
+
+I guess, I could run `git rebase -i` to squash commits to those files into one commit, thereby getting rid of the references to the old copies, but that approach feels awkward and error prone. Is anyone aware of a better way?

doc/forum/How_to_expire_old_versions_of_files_that_have_been_edited__63__/comment_1_dccf4dc4483d08e5e2936b2cadeafeaf._comment

@@ -0,0 +1,8 @@
+[[!comment format=mdwn
+ username="http://peter-simons.myopenid.com/"
+ ip="77.186.179.173"
+ subject="comment 1"
+ date="2012-02-09T18:53:00Z"
+ content="""
+Sorry for commmenting on my own question ... I think I just figured out that `git annex unused` *does* in fact do what I want. When I tried it, it just didn't show the obsolete versions of the files I edited because I hadn't yet synchronized all repositories, so that was why the obsolete versions were still considered used.
+"""]]

doc/forum/How_to_expire_old_versions_of_files_that_have_been_edited__63__/comment_2_5710294c1c8652c12b6df2233255a45e._comment

@@ -0,0 +1,8 @@
+[[!comment format=mdwn
+ username="http://joey.kitenet.net/"
+ nickname="joey"
+ subject="comment 2"
+ date="2012-02-09T19:42:28Z"
+ content="""
+Yes, contents are still considered used while tags or refs refer to them. Including remote tracking branches like `remotes/origin/master`
+"""]]

doc/tips/using_the_web_as_a_special_remote.mdwn

@@ -8,10 +8,10 @@ The web can be used as a [[special_remote|special_remotes]] too.
 Now the file is downloaded, and has been added to the annex like any other
 file. So it can be renamed, copied to other repositories, and so on.
 
-Note that git-annex assumes that, if the web site does not 404, the file is
-still present on the web, and this counts as one [[copy|copies]] of the
-file. So it will let you remove your last copy, trusting it can be
-downloaded again:
+Note that git-annex assumes that, if the web site does not 404, and has the
+right file size, the file is still present on the web, and this counts as
+one [[copy|copies]] of the file. So it will let you remove your last copy,
+trusting it can be downloaded again:
 
 	# git annex drop example.com_video.mpeg
 	drop example.com_video.mpeg (checking http://example.com/video.mpeg) ok